Research map / Vision-language and multimodal / Robot action and navigation
NavWM: A Unified Navigation World Model for Foresight-Driven Planning
Yanghong Mei, Longteng Guo, MingMing Yu, Guiyu Zhao, Xingjian He, Jing Liu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add latent world reasoning and world model to action prediction (ablation) | NavWM · Module ablation (RECON+HuRoN) | ATE: 0.513 -> 0.254 | Improved |
| add world model to action prediction (ablation) | NavWM · Module ablation (RECON+HuRoN) | ATE: 0.513 -> 0.338 | Improved |
| add latent world reasoning to AP + WM (ablation) | NavWM · Module ablation (RECON+HuRoN) | ATE: 0.338 -> 0.254 | Improved |
| add action prediction to WR + WM (ablation) | NavWM · Module ablation (RECON+HuRoN) | PSNR: 16.627 -> 17.622 | Improved |
| add latent world reasoning to AP + WM (ablation) | NavWM · Module ablation (RECON+HuRoN) | PSNR: 14.286 -> 17.622 | Improved |
| add second-stage fine-tuning on predicted trajectories (ablation) | NavWM world model · Training stage ablation | image reconstruction quality | Improved |
| use anchor-based multimodal head instead of GMM (with WM planning) (ablation) | NavWM · Trajectory head ablation (RECON+HuRoN) | ATE: 0.79 -> 0.36 | Improved |
| use anchor-based multimodal head instead of diffusion head (with WM planning) (ablation) | NavWM · Trajectory head ablation (RECON+HuRoN) | ATE: 0.58 -> 0.36 | Improved |
| use anchor-based head instead of GMM under raw sampling (ablation) | NavWM · Trajectory head ablation (RECON+HuRoN) | ATE: 0.88 -> 0.85 | No clear effect |
| use anchor-based head for trajectory diversity (ablation) | NavWM · Trajectory head ablation (RECON+HuRoN) | APD: 0.31 -> 1.49 | Improved |
| add world model planning to anchor-based trajectories (ablation) | NavWM · Trajectory head ablation (RECON+HuRoN) | ATE: 0.85 -> 0.36 | Improved |
| add world model planning to GMM trajectories (ablation) | GMM head · Trajectory head ablation (RECON+HuRoN) | ATE: 0.88 -> 0.79 | Improved |
Closest papers by meaning
- Towards Unified World Models for Visual Navigation via Memory-Augmented Planning and Foresight (ECCV 2026 · similarity 0.93)
- Navigation World Models (CVPR 2025 · similarity 0.91)
- Drift-Resistant Navigation World Model with Anchored Epipolar Guidance (NeurIPS 2026 · similarity 0.89)
- From Forecasting to Planning: Policy World Model for Collaborative State-Action Prediction (NeurIPS 2025 · similarity 0.88)
- NavGPT: Explicit Reasoning in Vision-and-Language Navigation with Large Language Models (AAAI 2024 · similarity 0.88)
- Predictive but Not Plannable: RC-aux for Latent World Models (NeurIPS 2026 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
